© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 81092
    Prev
    Next
    INDEX
    Explanations

    protocol buffer schema

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron reliably fires on tokens that commonly appear in file‐header comment blocks—e.g. comment delimiters (“//”), URLs (https, “.com”), dashes, quotes, and other metadata markers—so it’s detecting the start of license/documentation header comments.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    leski
    -0.75
    予算
    -0.72
    )」
    -0.71
    Als
    -0.70
    公众号
    -0.69
    gië
    -0.69
    ӣ
    -0.69
     alert
    -0.68
    EXTR
    -0.68
    isBlank
    -0.67
    POSITIVE LOGITS
    ㅛ
    0.91
     Uint
    0.89
    rpc
    0.89
    bears
    0.86
    carrying
    0.84
    repeated
    0.84
    般
    0.84
     wien
    0.83
    🍖
    0.82
    лое
    0.81
    Activations Density 0.007%

    No Known Activations